This is the rare piece that gives you the actual number instead of vibes: plan a sprint to about 80% and hold 20% back as an explicit interrupt buffer, because the interrupts are coming whether you budget for them or not. It also solves the part everyone gets wrong, who catches the fire: instead of yanking whoever is nearest off deep work, you rotate one person into an interrupt role for the cycle so the emergency lands on a designated buffer rather than blowing up three people's plans. For a small Indian team where the founder is often the one production issue away from losing a marquee client, that split is the difference between a bad Tuesday and a wrecked week.
How engineering teams handle unplanned work
From Boldstart Ventures by Anna Debenham 10 min read
- Plan to ~80% and reserve ~20% as a named buffer for interrupts, so unplanned work has a home instead of eating your priorities
- Rotate a single 'interrupt' owner per cycle so fires hit a designated buffer, not everyone's focus time at once
- Treat recurring interrupts as a signal: better QA, scheduled maintenance, and incident write-ups shrink next week's fire load
